这篇论文研究的是Single-cell RNA sequencing (scRNA-seq) denoising, 也就是单细胞RNA测序的降噪,由于数据扩增和数据丢失等问题,会干扰scRNA-seq的数据分析,因此需要有降噪技术用于稀疏的scRNA-seq数据,作者提出了一种deep count autoencoder network (DCA),通过negative binomial noise model with or without zero-...
A deep count autoencoder network to denoise scRNA-seq data and remove the dropout effect by taking the count structure, overdispersed nature and sparsity of the data into account using a deep autoencoder with zero-inflated negative binomial (ZINB) loss function. ...
so scalable denoising methods for increasingly large but sparse scRNA-seq data are needed. We propose a deep count autoencoder network (DCA) to denoise scRNA-seq datasets. DCA takes the count distribution, overdispersion and sparsity of the data into account using a negative binomial noise model ...
AEs Autoencoders ANNs Artificial Neural Networks BBKNN batch balanced k-nearest neighbors DL Deep Learning DNNs Deep neural networks DCA Deep Count AE network ERCC External RNA Controls Consortium HDST high-definition spatial transcriptomics FFNNs Feed Forward Neural Networks ML machine learning MMD Maxi...
Multiple statistical methods have been developed to impute and denoise scRNA-Seq data. Most of these methods rely on distribution assumptions on scRNA-Seq data matrix. For example, deep count autoencoder (DCA)9assumes negative binomial distribution with or without zero inflation, SAVER10assumes negati...
DCA is a deep learning-based method for denoising single-cell count matrices. DCA is implemented in Python and uses an autoencoder with a zero-inflated negative binomial (ZINB) loss function. For each gene, DCA computes gene-specific parameters of ZINB distribution, namely dropout, dispersion, ...
2.13.7. DCA -> code for 2022 paper: Deep Covariance Alignment for Domain Adaptive Remote Sensing Image Segmentation2.13.8. SCAttNet -> Semantic Segmentation Network with Spatial and Channel Attention Mechanism2.13.9. unetseg -> A set of classes and CLI tools for training a semantic segmentation...
[8]. More recently, deep learning-based approaches have been developed to overcome the scalability issue by conventional approaches. For example, scVI, scScope, and DCA use deep autoencoder (AE) to learn feature representation to recover dropouts and DeepImpute uses a deep neural network to learn...
作者在这里提出DCA(deep count autoencoder network)对scRNA-seq进行降噪;DCA使用binomial noise model with or without zero-inflation考虑到count distribution,overdispersionandsparsity of the data; 使用模拟数据和真实数据作者证明了DCA降噪可以提高各种各样传统的scRNA-seq分析,DCA在数据填充和速度上也比目前已有的方...
Autoencoders widely used for transcriptomics applications are shown to perform well on several tasks, like drug perturbation prediction [23] or dropout imputation [12]. Since the ordering of the genes in scRNA-seq count matrices is mostly arbitrary, fully connected layers are usually used in this...